IGME: Efficient Chained Method Ensemble for Transferable Semantic Segmentation Attacks
Quick Answer
The IGME method offers an efficient single-source approach for transferable semantic segmentation attacks, significantly reducing computational costs compared to ensemble attacks.
Quick Take
It utilizes a chained computation of differentiable components and an integrated-gradient-style stabilization heuristic, achieving competitive transferability on Pascal VOC and Cityscapes datasets with only one source model.
Key Points
- IGME reduces computational costs by using a single-source model for attacks.
- Chained computation allows sharing of expensive source-model gradient calculations.
- Integrated-gradient-style stabilization improves update stability during attacks.
- Achieves competitive transferability compared to single-source baselines.
- Evaluated on Pascal VOC and Cityscapes with CNN and transformer models.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Semantic segmentation models are vulnerable to transferable adversarial perturbations, yet evaluating transfer attacks on dense prediction models can be computationally expensive. Existing ensemble attacks often rely on multiple surrogate models, increasing the computation cost, even harder for segmentation. This paper studies an efficient single-source alternative for transferable attacks on semantic segmentation. We formulate transferable attack composition as a chained computation over differentiable attack components, allowing the expensive source-model gradient computation to be shared. To reduce the update instability introduced by chained composition, we further use an integrated-gradient-style path-averaged direction as an empirical stabilization heuristic. Experiments on Pascal VOC and Cityscapes evaluate the resulting transferability efficiency trade-off across CNN- and transformer-based segmentation models. IGME achieves competitive transferability compared with single-source baselines and favorable runtime compared with model-ensemble attacks, while requiring access to only one source model.
| Comments: | 8 pages, 3 figures |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| ACM classes: | I.2.10; I.4.6 |
| Cite as: | arXiv:2607.27465 [cs.CV] |
| (or arXiv:2607.27465v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.27465 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Mengqi He [view email]
[v1]
Wed, 29 Jul 2026 21:08:48 UTC (6,449 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.